Etched, an AI chip startup founded in 2022 by three Harvard dropouts, has secured a $300 million Series C funding round, elevating its valuation to $10.3 billion. This round was led by Sequoia, with participation from Andreessen Horowitz, SK Hynix, Jane Street, and Diffusion Capital, among others. Notable individual investors include Peter Thiel, Andrej Karpathy, Dylan Field, and Amjad Masad.
This latest valuation marks a significant increase from the $5 billion valuation achieved in December, following a $500 million funding round. In June, Etched announced the successful manufacturing of its proprietary chips, with initial systems undergoing client testing and $1 billion in orders already secured.
Etched’s inception coincided with skepticism about developing chips specifically tailored for AI models based on transformer technology—the architecture underpinning modern AI systems like ChatGPT and Claude. Despite initial doubts, the company has demonstrated that its systems can run a variety of AI models, including Mixture of Experts models like DeepSeek and Qwen, as well as non-transformer designs such as Mamba, which utilizes a state-space model architecture.
A key innovation from Etched is the development of two components designed to accelerate the inference process—the computation that occurs after a user submits a prompt. The company introduced a prefill chip that operates at a lower voltage, reducing heat generation and allowing for a higher transistor density. Additionally, Etched developed a new memory and interconnect technology called cluster scale memory, enabling multiple chips to share a fast, low-latency memory pool. These advancements aim to deliver higher speeds at reduced costs.
Etched’s rapid ascent and technological breakthroughs underscore the growing demand for specialized AI hardware capable of efficiently handling complex computations. As AI applications continue to expand, the company’s innovations position it as a formidable competitor in the AI chip market, challenging established players like Nvidia.
